| 00:11.49 | X-Rob | Holy crap, even on a freshly restarted system a fork is taking 5msec. That seems massive. |
| 00:11.53 | X-Rob | 2018-07-28 00:11:20] NOTICE[25617][C-00000001]: app.c:3030 ast_safe_fork: fork() call took 5ms |
| 00:21.39 | *** join/#asterisk-dev infobot (ibot@208.53.50.136) |
| 00:21.39 | *** topic/#asterisk-dev is Asterisk Development Discussion -=- http://www.asterisk.org/developers -=- Tier 2 and 3.14159265 support is in #asterisk -=- Check out our blog! blogs.asterisk.org -=- Follow on Twitter at @AsteriskDev |
| 00:48.43 | X-Rob | https://www.irccloud.com/pastebin/dNSWts8e/ubuntu.txt |
| 00:49.25 | X-Rob | https://www.irccloud.com/pastebin/lsPT478l/rhel7.txt |
| 00:50.12 | X-Rob | glibc has changed the way they implement fork() after 2.2.something, and it now uses clone() and a bunch of other stuff to make it faster. |
| 02:13.04 | X-Rob | Hah. Nope. It's gcc. |
| 02:13.13 | X-Rob | compile with a newer version of gcc, no forklag. |
| 02:24.39 | X-Rob | Yep, solved. Thanks file and gtjoseph for your help! |
| 02:24.40 | X-Rob | https://issues.asterisk.org/jira/browse/ASTERISK-27987?focusedCommentId=244311&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-244311 |
| 02:56.00 | X-Rob | .... maybe. |
| 02:56.05 | X-Rob | maybe it's being optimized out. |
| 03:24.39 | X-Rob | It is. *sob* |
| 04:21.59 | *** join/#asterisk-dev coreyfarrell (~coreyfarr@24-177-250-191.dhcp.nwtn.ct.charter.com) |
| 04:21.59 | *** mode/#asterisk-dev [+o coreyfarrell] by ChanServ |
| 07:13.15 | *** join/#asterisk-dev jkroon (~jkroon@165.16.204.164) |
| 09:53.35 | *** join/#asterisk-dev sh_smith (~sh_smith@cpe-76-174-26-91.socal.res.rr.com) |
| 10:52.37 | *** join/#asterisk-dev CELYA_1 (~Thunderbi@be.celya.fr) |
| 11:35.45 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d003ddfeb4c6a06811e.ipv6.abo.wanadoo.fr) |
| 11:47.12 | *** join/#asterisk-dev CELYA_1 (~Thunderbi@LFbn-1-5196-144.w90-105.abo.wanadoo.fr) |
| 13:32.39 | *** join/#asterisk-dev jkroon (~jkroon@165.16.204.174) |
| 13:40.24 | seanbright | well, not that it addresses the issue specifically, but fastagi is the way to go regardless |
| 13:52.09 | seanbright | i know very little about it, but maybe you could fork an agi "worker" process early on asterisk startup that has a smaller footprint and then *that* process could fork for each AGI call |
| 13:52.46 | seanbright | and then i guess you would forward the AGI req/resp to asterisk from there |
| 14:25.26 | *** join/#asterisk-dev jkroon (~jkroon@165.16.204.171) |
| 16:44.07 | *** join/#asterisk-dev CELYA_1 (~Thunderbi@2a01cb0583d31d009dd2a3780b41d8e9.ipv6.abo.wanadoo.fr) |
| 17:46.22 | *** join/#asterisk-dev tuxd00d (~tuxd00d@unaffiliated/tuxd00d) |
| 18:31.08 | *** join/#asterisk-dev cresl1n (uid299068@asterisk/libpri-and-libss7-expert/Cresl1n) |
| 18:31.08 | *** mode/#asterisk-dev [+o cresl1n] by ChanServ |
| 19:54.08 | *** join/#asterisk-dev tzafrir (~quassel@lemon.tzafrir.org.il) |
| 19:54.08 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |